引導式測驗 · 硬 等級
AI數位教育測驗
測驗學習目標、學習者隱私、可近性、教師評審、獨立解決問題和學習成果。
問題 1 的 8
What should an educational AI feature begin with?
預覽每個問題8 問題
當您參加測驗時,就會出現答案和解釋。
- What should an educational AI feature begin with?
- A learning objective and the decision it supports for a teacher or learner
- A target number of student messages
- The model’s most fluent subject
- A leaderboard position
- Why is interface activity alone insufficient evidence of learning?
- Activity cannot be measured
- Learning is identical to completion
- The model score already measures comprehension
- Clicks and time spent do not show that learners understood or can apply the material
- A generated translation sounds fluent but changes a concept. What does this illustrate?
- Translation quality is irrelevant to learning
- A fluent answer must be correct
- Fluency is not enough to establish accurate educational feedback
- Learners cannot notice translation errors
- When should teacher review be retained?
- Only when a student asks twice
- For high-impact assessment and explanations that need evidence
- Never, if the assistant cites a source
- Only for lessons with images
- What data practice supports student privacy?
- Send complete student histories to every tool
- Keep student data indefinitely
- Assume classroom use implies consent
- Document collection, retention, provider access, and age-appropriate safeguards
- Why assess the tool on varied devices and accessibility technologies?
- Access and feedback may differ across learner conditions
- Every device renders the same experience
- Accessibility tools only affect appearance
- A single average reveals all learner needs
- A lesson is intended to build independent problem-solving. What should the explanation avoid?
- Offering a question that checks understanding
- Hiding the reasoning process or final answer in a way that removes the intended practice
- Allowing a learner to ask for a hint
- Showing evidence behind an explanation
- Which outcome set gives a fuller view of an education tool?
- Number of prompts alone
- Model parameters and latency only
- Learning gains, completion, confidence, and support burden
- How often students open the page